Scanning of the Dopamine D1 and D5 ReceptorGenes by REF in Neuropsychiatric Patients Revealsa Novel Missense Change at a Highly ConservedAmino Acid

Jinong Feng,1 Janet L. Sobell,1 Leonard L. Heston,2 Edwin H. Cook, Jr.,3 David Goldman,4 andSteve S. Sommer5*1Division of Molecular Medicine, Beckman Research Institute, City of Hope National Medical Center,Duarte, California

2Department of Psychiatry, University of Washington, Seattle, Washington3Departments of Psychiatry and Pediatrics, University of Chicago, Chicago, Illinois4National Institute on Alcohol Abuse and Alcoholism, Rockville, Maryland5Division of Human Genetics, Department of Molecular Diagnosis, Beckman Research Institute, City of HopeNational Medical Center, Duarte, California

In previous analyses of schizophrenic pa-tients, multiple missense changes and onenonsense change were identified in the D5dopamine receptor (DRD5) gene, but no se-quence changes of likely functional signifi-cance were identified in the D1 dopaminereceptor (DRD1) gene. In the present study,we examined these genes in patients withcertain other neuropsychiatric disordersthat may be related to dopaminergic dys-regulation. The coding regions of the DRD1and DRD5 genes were examined in 25 and 25autistic patients, 25 and 28 attention deficithyperactivity disorder patients, and 51 and43 alcoholic patients, respectively. In addi-tion, the DRD5 gene was examined in 75schizophrenic patients to search for addi-tional variants affecting protein structureor expression (VAPSEs). These patientswere analyzed with REF (restriction endo-nuclease fingerprinting), a hybrid betweenSSCP and restriction endonuclease diges-tion, which allows the entire coding se-quence to be screened in one lane of a gel.Approximately 800 kb of genomic sequencewere examined. No sequence changes wereidentified in the DRD1 gene among the 101patient samples analyzed. Two sequence

changes were identified in the DRD5 geneamong the 171 patient samples. These in-cluded one previously identified silent poly-morphism at base pair 978 (P326P). Thechange was identified in patients from alldisease categories and from different ethnicbackgrounds. One novel missense change,L88F, occurred in transmembrane domainII at a highly conserved amino acid in alldopamine receptors as well as in a1- and b-adrenergic receptors. The mutation wasidentified in a Caucasian male patient withautism. Further analysis is necessary to de-termine if this missense change is associ-ated with a particular neuropsychiatricphenotype. Am. J. Med. Genet. (Neuropsy-chiatr. Genet.) 81:172–178, 1998.© 1998 Wiley-Liss, Inc.

KEY WORDS: dopamine receptor; schizo-phrenia; autism; attentiondeficit/hyperactivity disor-der; alcoholism

INTRODUCTIONCandidate gene association studies have been devel-

oped as an adjunct to linkage analysis in the search forgenes predisposing to multifactorial diseases, such asschizophrenia and other neuropsychiatric disorders[Sobell et al., 1992]. In one type of candidate genestudy, termed VAPSE-based analysis, gene regions oflikely functional significance are examined directly forvariants affecting protein structure or expression(VAPSEs). Once a VAPSE of likely functional signifi-cance is found in a candidate gene from a subsample ofaffected individuals, the prevalence of that allele is

Contract grant sponsor: NIMH; Contract grant numbers:MH44276, MH52223, KO MH01389.

*Correspondence to: Steve S. Sommer, Division of Human Ge-netics, Department of Molecular Diagnosis, Beckman ResearchInstitute, City of Hope National Medical Center, Duarte, CA91010.

Received 15 July 1997; Revised 13 November 1997

American Journal of Medical Genetics (Neuropsychiatric Genetics) 81:172–178 (1998)

© 1998 Wiley-Liss, Inc.

compared in a large group of unrelated patients andethnically similar controls to determine if a disease as-sociation exists. Parental controls, if available, offerthe best approach for removing ethnicity as a confound-ing variable [Schaid and Sommer, 1994; Falk and Ru-binstein, 1987].

The rate-limiting step in VAPSE-based candidategene studies is the examination of the regions of likelyfunctional significance in search of sequence changes.To increase the rate at which a DNA sequence may beexamined, while retaining high sensitivity and speci-ficity, our laboratory has developed a battery of meth-ods for genomic screening [Liu et al., 1996; Liu andSommer, 1995; Sarkar et al., 1992], including restric-tion endonuclease fingerprinting (REF) [Liu and Som-mer, 1995; Liu et al., 1997]. REF allows the detection ofsingle base pair changes in a large segment of ampli-fied DNA (1–2 kb) by the creation of overlapping re-striction fragments that, when electrophoresedthrough a nondenaturing gel, can exhibit migrationaldifferences due to underlying conformational changes.Thus, a particular sequence change can be detected asan altered electrophoretic pattern in multiple seg-ments. Furthermore, if a sequence change alters a re-striction site, these changes may be readily identifiedas missing or added segments.

In the present study, REF was used to examine thedopamine D1 and D5 receptor (DRD1 and DRD5) genesas candidates for neuropsychiatric disease. DRD1 andDRD5 belong to a subfamily of dopamine receptorsknown to stimulate the production of adenylyl cyclasevia G-protein coupling, thereby activating cyclic AMP(c-AMP)-dependent protein kinases. Most availableagonists and antagonists of DRD1 interact similarlywith DRD5. The amino-acid sequence identity of DRD1and DRD5 in their characteristic seven transmem-brane domains is 80%, with an overall amino-acid ho-mology of 50% [Sunahara et al., 1991]. The proteinproducts are of average length, being 477 amino acidsfor D5 [Sunahara et al., 1991] and 446 residues for D1[Sunahara et al., 1990]. The coding region of both genesis uninterrupted. The genes have been localized tochromosomes 4p15.1–15.3 (DRD5) [Sherrington et al.,1993] and 5q35.1 (DRD1) [Grandy et al., 1990].

Involvement of dopaminergic systems has been hy-pothesized in a variety of psychiatric disorders. For ex-ample, dopaminergic activity in the striatum and lim-bic system has been suggested to underlie the develop-ment of substance dependence in animal models ofaddictive behavior (e.g., alcoholism) [Tiihonen et al.,1995]. Imbalances in noradrenergic and dopaminergicsystems may be associated with particular symptomsof attention deficit/hyperactivity disorder (ADHD)[Malone et al., 1994], and DRD1 nullizygous mice arehyperactive [Xu et al., 1994]. Selective destruction ofdopamine neurons by the injection of 6-hydroxydopa-mine in the neonatal rat brain has been shown to resultin behavioral problems, such as hyperactivity andlearning deficit symptoms [Garfinkel and Wender,1989]. Excesses of dopaminergic activity in the basalganglia have been associated with stereotyped repeti-tive behavior in experimental animals, while repetitive

behaviors have been associated with deficient activa-tion produced by frontal lobe lesioning [Ridley, 1994].

In previous analyses in schizophrenic patients, mul-tiple missense changes and one nonsense change wereidentified in DRD5 [Sobell et al., 1995], but no se-quence changes of likely functional significance wereidentified in DRD1 [Liu et al., 1995]. The purpose of thepresent study was to search for additional DRD1 orDRD5 sequence changes in an extended sample ofschizophrenic patients, as well as in patients withother neuropsychiatric disorders (alcoholism, attentiondeficit/hyperactivity disorder, or autism), that may berelated to dopaminergic dysregulation. An additionalpurpose was to demonstrate further the feasibility ofusing REF as a highly sensitive and specific screeningtest.

MATERIALS AND METHODSPatient Samples

All schizophrenic patients met the criteria for dis-ease as defined by the Diagnostic and StatisticalManual III, Revised (DSM-III-R), as described previ-ously [Sobell et al., 1993]. The majority of patients wereascertained through state mental institutions in Min-nesota, Washington, and Oregon. Unrelated alcoholicpatients of Finnish ethnicity were ascertained throughcollaborative efforts involving the National Institute onAlcohol Abuse and Alcoholism (NIAAA) and the Uni-versity of Helsinki, Finland (supervised by D.G.). TheFinnish alcoholics were a criminal-offender populationand were diagnosed by DSM-III criteria. The alcoholic-offender population is highly enriched for antisocialpersonality disorder and intermittent explosive disor-der. Southwestern Native American patients were as-certained through the NIAAA (by D.G.). The South-western American Indians were interviewed with theSchedule for Affective Disorders and Schizophrenia(SADS-L) and diagnosed in blind fashion using DMS-III-R criteria. None of the Southwestern Indian alco-holics were first-degree relatives and, although thesesubjects were all members of the same very large pedi-gree, their degree of relationship was equal to or lessthan the average degree of relationship for the popula-tion from which they were derived [Goldman et al.,1992, 1997]. ADHD patients were ascertained at theUniversity of Chicago [Cook et al., 1995] using the fol-lowing criteria: 1) child or adolescent with a DSM-III-Rdiagnosis of ADHD made in a consensus diagnosticconference in which a child psychologist, child psychia-trist, and a developmental pediatrician presented find-ings from each of their evaluations; and 2) consent toparticipate by both parents and child. Patients withautistic disorder were also ascertained at the Univer-sity of Chicago if the following criteria were met: 1)child or adolescent with a mental age of at least 18months and intelligence quotient greater than 35; 2)DMS-IV diagnosis of autistic disorder by a child psy-chologist and child psychiatrist; 3) diagnosis of autisticdisorder by the Autism Diagnostic Interview-Revised(ADI) [Lord et al., 1994]; 4) absence of specific identi-fied etiology based on history and physical examina-tion; and 5) consent to participate from both parents

DRD1/DRD5 and Neuropsychiatric Diseases 173

and child. Selected demographic characteristics of allpatients are displayed in Table I.

Patient 1242 With Autism and aMissense Mutation

There were 2 paternal cousins (in different sibships)with speech and language problems. The patient wasnot severe, had sleep problems which resolved, and hadmild comorbid attention deficit hyperactivity disorderfor which pemoline (Cylert) and methylphenidate wereprescribed sequentially with relatively little benefit.He did not have other comorbid features, specifically,neither seizures nor increased irritability.

When he was 4 years old, his ADI scores were 21(cutoff, 10), 13 (cutoff, 8), and 6 (cutoff, 3) in social,communicative, and restricted and repetitive interestsdomains, respectively. At age 6, his Differential AbilityScales (DAS) Verbal Standard Score was 92 and hisNonverbal Reasoning Score was 105. His receptive lan-guage, as measured by the Peabody Picture VocabularyTests-Revised, was average (standard score, 90).

Medical evaluation revealed a boy at the 61st centileof height and 80th centile of weight. He was not dys-morphic and his overall physical examination was nor-mal, except for his developmental problems. Labora-tory evaluation included normal brain-stem auditory-evoked response, normal urinary amino-acid andorganic-acid analysis, and normal electroencephalo-gram, with 46XY karyotype, and negative fragile Xchromosomal testing. His thryoxine level as slightlyincreased (13.3, with a reference range of 6.0–12.0 mcg/dl).

Laboratory Methods

DNA amplification. DNA was extracted as previ-ously described [Gustafson et al., 1987]. For both theDRD1 and DRD5 genes, an initial amplification byPCR was performed in a total volume of 25 ml with 10mM Tris-HCl, pH 8.3, 50 mM KCl, 1.5 mM MgCl2, 200mM of each deoxyribonucleoside triphosphate, and 0.1mM of primers specific for DRD1, i.e., DRD1 (Hs)-58UTR-(−29)-27D: 58GAGCCCCTGATGTGCTTT-CTCTTAGGA38; DRD1 (Hs)-38UTR-(1413)-27U:58TTAATAGCAAGCCCCAGAGCAATCTCC38, or ofprimers specific for DRD5, i.e., DRD5 (Hs)-58UTR-(−74)-32D: 58CCGATGGGGCTGCCTGGGGGTCGCAGGG-CTGA38; DRD5 (Hs)-38UTR-(1503)-32U: 58GCGTGTG-

TGTGCGTGCTTGTCAGTGTCTGTGC38. The informa-tive primer names and number systems for the geneshave been previously described [Grandy et al., 1991;Dearry et al., 1990; Stoflet et al., 1988]. 1 U of Ampli-Taq (Perkin-Elmer, Norwalk, CT) and 2 ng of genomicDNA were added (denaturation at 95°C for 15 sec, an-nealing at 65°C for 30 sec, and elongation at 72°C for 3min for a total of 15 cycles with the Perkin-Elmer Ge-neAmp PCR System 9600). A second nested round ofPCR was performed using internal primers for DRD1,i.e., DRD1 (Hs)-58UTR-(−7)-30D: 58TAGGAAGATGAG-GACTCTGAACACCTCTGC38; DRD1 (Hs)-38UTR-(1361)-28U: 58GGCAGGATTCATCTGCGAGTTCAG-GTTG38, and internal primers for DRD5, i.e., DRD5(Hs)-58UTR-(−7)-22D: 58GCCCGAAATGCTGCCGC-CAGGC38; DRD5 (Hs)-38UTR-(1450)-32U: 58GTTTCT-TAATGCAGTTTAATGGAATCCATTC38, using 1 ml ofthe first-round product. This was added to a volume of75 ml with the same reagents as above. PCR cycleswere conducted at the same temperatures and timesfor a total of 30 cycles and a final elongation for 10 minat 72°C. The nested amplification products were puri-fied using Microcon-100 columns (Amicon, Beverly,MA). Product sizes were 1,426 bp for DRD1 and 1,511bp for DRD5.

Overview of REF. REF is a hybrid between SSCPand restriction endonculease digestion in which virtu-ally all mutations in a 1–2-kb segment can be detected[Liu and Sommer, 1995; Liu et al., 1997; Fujimura etal., 1997]. If brief, the region of interest is amplified byPCR, and divided into multiple aliquots. Each aliquotis digested by a different set of endonucleases. The ali-quots are combined, end-labeled, denatured, and elec-trophoresed on a nondenaturing gel. If six different en-donuclease digestions are performed, 12 single-stranded segments on the REF fingerprint will containthat variant. If only one of those segments displays analtered mobility, the variant will be detected. In addi-tion to the SSCP effect, approximately one quarter ofthe mutations will be detected because they create oreliminate a restriction site. Our revised protocol fol-lows.

Restriction endonuclease digestions. The re-striction enzyme digestions were performed as de-scribed previously [Liu and Sommer, 1995]. Selectionof the enzyme battery was aided by computer algo-rithms developed in our laboratory [Scaringe et al., un-

TABLE I. Distribution of DRD5 and DRD1 Patients by Diagnosis, Race/Ethnicity, and Gender (M/F)

Diagnosis Schizophrenia, DRD5

Autism ADHD Alcoholism

DRD1 DRD5 DRD1 DRD5 DRD1 DRD5

Race/ethnicityCaucasian 21/3 21/3 17/6 19/7

Western European 41/19Northern European 17/0a 16/0a

Hispanic 1/0 1/0African-American 13/2 0/1 0/1Native American 21/13b 13/14b

Asian 1/0 1/0Total 75 25 25 25 28 51 43

aFinnish.bSouthwestern American Indian.

174 Feng et al.

published findings]. In brief, 100 ng of purified DNAwere digested in separate tubes by four groups of en-donucleases (AvaII/BbsI, BbvI, DrdI/MslI, and MboIfor DRD1, and AluI, BsaHI, MslI, and PleI/BbsI forDRD5). When two enzymes were used, the digestionswere performed at one time in the same tube. All en-zymes and buffers were purchased from New EnglandBiolabs (Beverly, MA). Digestion reactions were incu-bated at 37°C for 6–8 hr, followed by enzyme inactiva-tion by heating at 80°C for 30 min.

End-labeling, electrophoresis, and sequenc-ing. The digestion reaction products were combinedand incubated with 0.5 U of calf intestinal alkalinephosphatase (CIAP) (Life Technologies, Gaithersburg,MD) at 37°C for 2 hr to remove the phosphate group atthe 58 end of the digested DNA fragments, and thenheated to 80°C for 30 min. Four nanograms of digestedDNA products were 58 end-labeled with [g-33P] ATP(Amersham, Arlington Heights, IL) and 1 U of T4 poly-nucleotide kinase (Promega, Madison, WI) in 2 ml re-actions containing 50 mM Tris-HCl, pH 7.4, 10 mMMgCl2, and 5 mM DTT. Incubation was at 37°C for 30min. The end-labeled product (1.5 ml) was electropho-resed at room temperature on a 0.5 × MDE™ gel with2% urea (0.33 mol/l), using the Model SE 1500 PokerFace™ sequencing apparatus (Hoefer Scientific, SanFrancisco, CA) with a 50 mM Tris-borate (pH 8.3)buffer at 15 W constant power. Autoradiographs of therestriction endonuclease fingerprints were interpretedas previously described [Liu and Sommer, 1995].Samples with abnormal or suspicious banding patternswere cycle-sequenced with the Perkin-Elmer GeneAmpPCR System 9600, with cycle parameters of denatur-ation at 95°C for 15 sec, annealing at 55°C for 30 sec,and elongation at 72°C for 1 min for a total of 30 cycles.Sequence analysis was performed with cycle sequenc-ing or GAWTS (genomic amplification with transcriptsequencing) [Stoflet et al., 1988].

Blinded analysis. Included among the DRD5 pa-tient samples were 8 schizophrenic individuals previ-ously shown to have eight different sequence changesin the coding region of the D5 dopamine receptor, rang-ing from 66–1,358 bp [Sobell et al., 1995]. A total of 14different sequence change events occurred in thesamples: 5 of the 8 unique changes (C806T, C989A,C1005A, A1051G, and C1358G) occurred in only 1 in-

dividual each, while one change (T978C) occurred in 5individuals, and two (G66A and G990A) occurred in 2individuals each. Blinded analysis was performed withthese samples, as the laboratory investigator (J.F.) didnot kown which samples were controls, how many posi-tive controls were included, or the types of sequencechanges represented in the positive samples.

RESULTSSequence Changes

For the DRD1 gene, no sequence changes were iden-tified among the 102 patient samples analyzed. Twosequence changes in the DRD5 gene were identified byREF. These included one previously known silent poly-morphism at base pair 978 (P326P). The change wasidentified in patients from all disease categories andfrom different ethnic backgrounds (Table II). One novelmissense change, L88F, which results from a C→Ttransition at base pair 262 (CTT→TTT), was identifiedin patient 1242, a Caucasian male with autism. In ad-dition, a C→A transversion at base pair 989 resulted ina proline-to-glutamine missense change at amino acid330 (P330Q). This previously described sequencechange was identified serendipitously as a result of di-rect sequencing, but was not evident on the REF gel.

DISCUSSION

This study represents the first step in VAPSE-basedcandidate gene association studies [Sobell et al., 1992].This approach seeks first to delineate sequencechanges of likely functional significance in candidategenes, and then to test aberrant alleles for disease as-sociation. VAPSE-based strategies have been devel-oped as an alternative approach to linkage-basedanalysis and to linkage disequilibrium-based ap-proaches in the study of multifactorial disorders [Sobelland Sommer, 1997; Sobell et al., 1992].

The rate-limiting step in the conduct of VAPSE-based analyses is the identification of sequencechanges of likely functional significance. Recently, hy-brids between SSCP and other methods have been de-veloped for efficiently scanning candidate genes [Liuand Sommer, 1995; Grompe, 1993; Sarkar et al., 1992].Restriction endonuclease fingerprinting is one suchmethod; REF was shown by blinded analysis to be

TABLE II. Frequency by Race of T978C Polymorphism Identified Among DRD5Alleles in Patients With Neuropsychiatric Disease*

Disease Race/ethnicity Total alleles C allele frequency

Schizophrenia Caucasiana 120 25.8%African American 30 33.3%

Autism Caucasiana 46 30.4%ADHD Caucasiana 52 26.9%

African American 2 0.0%Alcoholism Caucasian (Finnish) 32 21.9%

Native Americanb 54 40.7%Total Caucasian 218 27.1%

Caucasian (Finnish) 32 21.9%African American 32 31.2%Native American 54 40.7%

*Excludes Hispanic, Asian, and Asian Indian alleles due to small numbers.aWestern European descent.bSouthwestern American Indian.

DRD1/DRD5 and Neuropsychiatric Diseases 175

highly sensitive for the detection of single base pairmutations of all types [Liu and Sommer, 1995]. REFwas utilized in the present analysis to examine the D1and D5 dopamine receptor genes in a large group ofpatients with neuropsychiatric disease.

Sensitivity of REF

By decreasing the number of restriction endonucle-ases used from six to four and including samples fromindividuals with known mutations (see Materials andMethods), further data on the sensitivity of REF inblinded analyses were obtained. One of 14 known mu-tations was missed when REF was performed on a1.5kb segment with only four groups of restriction en-donucleases. Based on past blinded analyses in whichall mutations were detected, five enzymes are recom-mended when REF is performed on a 1kb segment, andsix enzymes are recommended for a 1.3–2.3kb segment[Liu and Sommer, 1995]. The recent development of‘‘REF Select’’ software (available on request) facilitatesthe selection of appropriate restriction endonucleasegroups, which are important to ensure that virtually allmutations are detected with the above-mentionednumber of enzyme groups [Scaringe et al., unpublishedobservations].

Previously Observed Sequence Changes

REF analysis of the DRD1 gene in 101 samples re-vealed no sequence changes. Several sequence changespreviously have been reported, including G(−94)A [Chi-con et al., 1994]; A(−48)G [Liu et al., 1995; Ohara et al.,1993]; A90G [Ohara et al., 1993]; G198A [Liu et al.,1995]; G1263A [Liu et al., 1995; Chicon et al., 1994],and T1403C [Chicon et al., 1994]. As the current analy-sis screened a PCR product from −7–1361, polymor-phisms at basepairs −94, −48, and 1403 would not bedetectable. The G198A change, reported in our previ-ous analysis of the DRD1 gene in schizophrenics, wasobserved only in Asians, while the G1263A polymor-phism was found in 2.8% of Caucasians of WesternEuropean descent (Finns were not included). Only oneAsian patient was included in the present DRD1 analy-sis. Although 128 Caucasian alleles were included

(Table I), only 94 were from non-Finnish Caucasians.The failure to observe the G1263A polymorphism when2.6 were expected is not significant.

For the DRD5 gene, 171 patients were analyzed, andtwo previously identified sequence changes were ob-served: a polymorphism (T978C) and a known mis-sense change in a nonconserved amino acid (P330Q). Inaddition, a novel missense change, L88F, was identi-fied within transmembrane domain II of the protein ina Caucasian male patient with autism. A comparison ofthe DRD5 protein sequence between receptor subtypesand species showed that this amino acid is highly con-served (Fig. 1).

TABLE III. Summary of VAPSEs Identified in Selected Catecholamine System and Other Genes in Patients With Schizophreniaor Other Mental Illnesses

Genea Patientsb

Uniquesequence

scanned (bp)

Totalsequence

scanned (kb)Screeningmethodc

No. ofVAPSEs

No. ofnon-VAPSEs References

DRD1 156 (76) 1,400 643 ddF; REF 0 3 Liu et al., 1995; this studyDRD2 14 3,464 97 GAWTS 0 3 Sarkar et al., 1991DRD5 153 (96) 1,573 743 ddF; REF 6 4 Sobell et al., 1995; this studya2AAR 93 (113) 1,584 653 REF 4 2 Feng et al., 1998b2AR 95 1,047 214 Bi-ddF 0 6 Feng et al., unpublished dataMAO-B 100 2,350 235 ddF; Bi-ddF; GAWTS 0 8 Sobell et al., 1997COMT 100 1,800 360 ddF; GAWTS 2 6 Sobell et al., unpublished dataPENK-A 150 1,820 546 ddF; SSCP 1 6 Mikesell et al., 1996APP 212 392 166 GAWTS 0 0 Arnholt et al., 1993Total 1,358 15,430 3,657 13 38

aa2AAR, a2 adrenergic receptor gene, subtype A; b2AR, b2 adrenergic receptor gene; MAO-B, monoamine oxidase B gene; COMT, catechol-o-methyltransferase gene; PENK-A, proenkephalin A gene; APP, amyloid precusor protein gene.bOther mental illness patients indicated in parentheses.cddF, dideoxy fingerprinting; SSCP, single-strand conformation polymorphism; Bi-ddF, bidirectional dideoxy fingerprinting.

Fig. 1. Amino-acid alignment of DRD5 with other dopamine receptorsindicates that L88 is completely conserved in this gene family. An addi-tional alignment with the adrenergic receptor gene family revealed thatL88 was present in the adrenergic receptors a1A, B, and C, and a adren-ergic receptors b, 1, 2, and 3, and conservatively changed to I in the ad-renergic receptors a 2A, B, and C (data not shown).

176 Feng et al.

Based on the high level of evolutionary conservation,the missense change is highly likely to be of functionalsignificance. A multitude of data suggests that thelevel of conservation is a good indicator of this likeli-hood; for example, analyses of the relationship betweenmissense mutations causing hemophilia B and the ex-tent to which residues in factor IX are conserved duringevolution indicate that almost all missense mutationsthat occur at residues conserved in the factor IX genefamily (more than 2 billion years of evolutionary diver-gence) are functionally deleterious [Bottema et al.,1991]. This same relationship has been found for othergenes in which many mutations have been described(e.g., p53, phenylalanine hydroxylase, and the cysticfibrosis transmembrane conductance regulator). Fu-ture studies with large samples are needed to deter-mine the possible association of the L88F VAPSE withneuropsychiatric disease.

If no association is found between this VAPSE andneuropsychiatric disease, additional studies may beconducted to determine if the aberrant allele is associ-ated with other disorders or clinical traits. For theseanalyses, a large number (preferably thousands) ofDNA samples may be screened to identify a subset ofindividuals with the aberrant allele. Analyses of medi-cal history between individuals with the VAPSE andthose homozygous for the wild-type allele may be con-ducted.

Based on studies in Drosophila, recessive lethal mu-tations reduce fitness in heterozygotes by an average of3% [Crow, 1993]. Most deaths due to the mutation occurin the heterozygotes (rather than the homozygotes) be-cause heterozygotes are much more frequent in thepopulation. These ‘‘genetic deaths’’ in heterozygotes oc-cur before the end of the reproductive period. However,additional morbidity and mortality are likely to occurafter the reproductive period. If these data extrapolateto humans, one would expect that mutations at highlyconserved residues would often predispose to diseasesor traits in heterozygotes, especially at higher ages.

Frequency of VAPSEs

Genomic sequence (3.6. megabases) has been ana-lyzed in a total of nine genes. When an average of 151patients was scanned within an average of 1,714 bp ofregions of likely functional significance, an average of1.4 VAPSEs was found. Of these, 46% are very likely tobe of functional significance (e.g., a nonsense mutationor a missense mutation at an amino acid that has beenconserved during more than 2 billion years of evolu-tionary divergence). The remaining VAPSEs are mis-sense changes that represent a mixture of functionallyimportant changes and neutral polymorphisms.

Two classes of genes emerge: those genes with noVAPSEs despite extensive scanning for mutations (e.g.,DRD1), and those genes with VAPSEs (sometimes withfour or more VAPSEs).

For every VAPSE found within regions of likely func-tional significance, there were approximately threenon-VAPSE variants, mostly silent changes in the cod-ing region. Within the coding sequence, there were twoVAPSEs for every silent change (13 VAPSEs vs. 25silent changes).

Candidate Genes Analyzed

To partially examine the dopamine hypothesis ofschizophrenia, our laboratory has examined genes inthe receptor and degradative pathways. These have in-cluded the D1, D2, and D5 dopamine receptor genes, aswell as catechol-o-methyltransferase and monoamineoxidase B. Other noncatecholamine genes have alsobeen examined, including dystrophin, proenkephalinA, adrenergic receptors, and the amyloid precursor pro-tein gene. Within the power of our sample size, none ofthe VAPSEs correlated with disease in case-controlanalyses. Our sample sizes were often adequate to de-tect attributable risks on the order of 4%. Although theresults of association studies have been negative, thedata are still useful in ruling out certain genes as com-mon causes of schizophrenia.

It should be noted that the strength of the VAPSE-based association study approach does not lie in itsapplication by any one group of investigators, any morethan the usefulness of the linkage approach in multi-factorial diseases can be judged by the results of anysingle group. Rather, the VAPSE-based approach willhave its greatest usefulness when applied by many in-dependent investigative groups in a variety of patientpopulations. Ultimately, as technological advancesmake direct gene examinations efficient, the identifi-cation of sequence changes of likely functional signifi-cance will be commonplace.

ACKNOWLEDGMENTS

We thank Mark Stein, Catherine Lord, Marrea Win-ega, and Bennett Leventhal for diagnostic assistance,and Shuya Yan for expert technical assistance. Thiswork was partially supported by grants MH44276,MH52223,and KO2 MH01389 from NIMH.

REFERENCES

Arnholt JC, Sobell JL, Heston LL, Sommer SS (1993): Amyloid precursorprotein (APP) gene mutations and schizophrenia. Biol Psychiatry 34:739–740.

Bottema CDK, Ketterling RP, Ii S, Yoon H, Phillips JA III, Sommer SS(1991): Missense mutations and evolutionary conservation of aminoacids: Evidence that many of the amino acids in factor IX function as‘‘spacer’’ elements. Am J Hum Genet 49:820–838.

Chicon S, Nothen MM, Erdmann J, Propping P (1994): Detection of fourpolymorphic sites in the human dopamine D1 receptor gene (DRD1).Hum Mol Genet 3:209.

Cook EH, Stein MA, Krasowski MD, Cox NJ, Olkon DM, Kieffer JE, Lev-enthal BL (1995): Association of attention-deficit disorder and the do-pamine transporter gene. Am J Hum Genet 56:993–998.

Crow JF (1993): How much do we know about spontaneous human muta-tion rates? Environ Mol Mutagen 21:122–129.

Dearry A, Gingrich JA, Falardeau P, Fremeau RT, Bates MD, Caron MG(1990): Molecular cloning and expression of the gene for a human D1dopamine receptor. Nature 347:72–75.

Falk CT, Rubinstein P (1987): Haplotype relative risks: An easy, reliableway to construct a proper control sample for risk calculations. AnnHum Genet 51:227–233.

Feng J, Sobell JL, Heston LL, Goldman D, Cook EH Jr, Gelernter J, Kran-zler H, Sommer SS (1998): Variants in the a2A adrenergic receptor genein psychiatric patients.

Fujimura FK, Liu Q, Feng J, Sommer SS (199): Detection of mutation bysingle-strand confirmation polymorphism (SSCP) analysis and SSCP-hybrid methods. Current Protocols in Human Genetics 7.4.1–7.4.23.

Garfinkel BD, Wender PH (1989): Attention-deficit hyperactivity disorder.

DRD1/DRD5 and Neuropsychiatric Diseases 177

In Kaplan HI, Sadock BJ (eds): ‘‘Comprehensive Textbook of Psychia-try.’’ Baltimore: Williams and Wilkins, p 1831.

Goldman D, Dean M, Brown GL, Bolos AM, Tokola R, Virkkunen M, Lin-noila M (1992): D2 dopamine receptor genotype and cerebrospinal fluidhomovanillic acid, 5-hydroxyindoleacetic acid and 3-methoxy-4hydroxyphenylphenylglycol in Finnish and American alcoholics. ActaPsychiatr Scand 86:351–357.

Goldman D, Urbanek M, Guenther D, Robin R, Long JC (1997): Linkageand association of a functional DRD2 variant [Ser311Cys] and DRD2markers to alcoholism, substance abuse, and schizophrenia in South-western American Indians. Am J Med Genet 74:386–394.

Grandy DK, Zhou QY, Allen L, Litt R, Magenis RE, Civelli O, Litt M (1990):A human D1 dopamine receptor gene is located on chromosome 5 atq35.1 and identifies an EcoRI RFLP. Am J Hum Genet 47:828–834.

Grandy DK, Zhang Y, Bouvier C, Zhou Q, Johnson RA, Allen L, Buck K,Bunzow JR, Salon J, Civelli O (1991): Multiple human D5 dopaminereceptor genes: A functional receptor and two pseudogenes. Proc NatlAcad Sci USA 88:9175–9179.

Grompe M (1993): The rapid detection of unknown mutations in nucleicacids. Nat Genet 5:111–117.

Gustafson S, Proper JA, Bowie EJW, Sommer SS (1987): Parameters af-fecting the yield of DNA from human blood. Anal Biochem 165:294–299.

Liu Q, Sommer SS (1995): Restriction endonuclease fingerprinting (REF):A sensitive method for screening mutations in long, contiguous seg-ments of DNA. Biotechniques 18:470–477.

Liu Q, Sobell JL, Sommer SS (1995): Screening the dopamine D1 receptorgene in 131 schizophrenics and eight alcoholics: Identification of poly-morphisms but lack of functionally significant sequence changes. Am JMed Genet 60:165–171.

Liu Q, Feng J, Sommer SS (1996): Bi-directional dideoxy fingerprinting(Bi-ddF): A rapid method for quantitative detection of mutations ingenomic regions of 300–600 bp. Hum Mol Genet 5:107–114.

Liu Q, Feng J, Sommer SS (1997): In a blinded analysis. Restriction En-donuclease Fingerprinting detects all the mutations in a 1.9kb seg-ment. Biotechniques 23:836–839.

Lord C, Rutter M, LeCouteur A (1994): Autism Diagnostic Interview-Revised: A revised version of a diagnostic interview for caregivers ofindividuals with possible pervasive developmental disorders. J AutismDev Disord 24:659–685.

Malone MA, Kershner JR, Swanson JM (1994): Hemispheric processingand methylphenidate effects in attention-deficit hyperactivity disorder.J Child Neurol 9:181–189.

Mikesell MJ, Sobell JL, Sommer SS, McMurray CT (1996): Identification ofa missence change and polymorphisms in the proenkephalin A gene ofschizophrenic patients. Am J Med Genet 67:459–467.

Ohara K, Ulpian C, Seeman P, Sunahara RK, Van Tol HH, Niznik HB(1993): Schizophrenia: Dopamine D1 receptor sequence is normal, buthas DNA polymorphisms. Neuropsychopharmacology 8:131–135.

Ridley RM (1994): The psychology of perservative and stereotyped behav-iour. Prog Neurobiol 44:221–231.

Sarker G, Kapelner S, Grandy DK, Marchionni M, Civelli O, Sobell J,Heston L, Sommer SS (1991): Direct sequencing of the dopamine D2receptor (DRD2) in schizophrenics reveals three polymorphisms but nostructural change in the receptor. Genomics 11:8–14.

Sarkar G, Yoon H, Sommer SS (1992): Dideoxy fingerprinting (ddF): Arapid and efficient screen for the presence of mutations. Genomics 13:441–443.

Schaid DJ, Sommer SS (1994): Comparison of statistics for candidate-geneassociation studies with cases and parents. Am J Hum Genet 55:402–409.

Sherrington R, Mankoo B, Attwood J, Kalsi G, Curtis D, Buetow KH, PoveyS, Gurling H (1993): Cloning of the human dopamine D5 receptor geneand identification of a highly polymorphic microsatellite for the DRD5locus that shows tight linkage to the chromosome 4p reference markerRAF1P1. Genomics 18:423–425.

Sobell JL, Sommer SS (1997) Genetics. In Tasman A, Kay J, Lieberman JA(eds): ‘‘Psychiatry.’’ Philadelphia: W.B. Saunders, pp 182–209.

Sobell JL, Heston LL, Sommer SS (1992): Delineation of genetic predispo-sition to multifactorial disease: A general approach on the threshold offeasibility. Genomics 12:1–6.

Sobell JL, Heston LL, Sommer SS (1993): Novel association approach fordetermining the genetic predisposition to schizophrenia: Case-controlresource and testing of the first candidate gene. Am J Med Genet 48:28–35.

Sobell JL, Lind TJ, Sigurdson DC, Zald DH, Snitz BE, Grove WW, HestonLL, Sommer SS (1995): The D5 dopamine receptor gene in schizophre-nia: Identification of a nonsense change and multiple missense changesbut lack of association with disease. Hum Mol Genet 4:507–514.

Sobell JL, Lind TJ, Hebrink DD, Heston LL, Sommer SS (1997): Screeningthe monoamine oxidase B gene in 100 male schizophrenics: A cluster ofpolymorphisms in African-Americans but lack of functionally signifi-cant sequence changes. Am J Med Genet (in press).

Stoflet ES, Koeberl DD, Sarkar G, Sommer SS (1988): Genomic amplifica-tion with transcript sequencing. Science 239:491–494.

Sunahara RK, Niznik HB, Weiner DM, Stormann TM, Brann MR,Kennedy JL, Gelernter JE, Rozmahel R, Yang YL, Israel Y, et al.(1990): Human dopamine D1 receptor encoded by an intronless gene onchromosome 5. Nature 347:80–83.

Sunahara RK, Guan H, O’Dowd BF, Seeman P, Laurier LG, Ng G, GeorgeSR, Torchia J, Van Tol HHM, Niznik HB (1991): Cloning of the gene fora human dopamine D5 receptor with higher affinity for dopamine thanD1. Nature 350:614–619.

Tiihonen J, Kuikka J, Bergstrom K, Hakola P, Karhu J, Byynanen OP,Fohr J (1995): Altered striatal dopamine re-uptake site densities inhabitually violent and non-violent alcoholics. Nat Med 1:654–657.

Xu M, Moratalla R, Gold LH, Hiroi N, Koob GF, Graybiel AM, Tonegawa S(1994): Dopamine D1 receptor mutant mice are deficient in striatalexpression of dynorphin and in dopamine-mediated behavioral re-sponses. Cell 79:729–742.

178 Feng et al.